Thread by @IDrawCharts

in a thread-writing mood (aka I'm bored and definitely not doing drugs)

So let's talk about volatility:

-what it is
-how it's calculated
-how it might be useful to point & click traders

What it is: in plain terms, volatility is a measure of how much a trading pair's price bounces around the benchmark value, which is usually set to the average closing price.

A trading pair that spikes and drops by a lot around its average price is considered volatile.

To calculate volatility we'll start with a price average. For this example we'll use a 3D average price (calculated from closing price) of BTC perps.

From October 6 through November 8, the average closing price was $64,189.

Now, we'll find the furthest traded price from $64,189 that occurred during each of the 3 days in our sample.

11/06: the low was 60,100 (- 4,089 from average)
11/07: the low was 61,394 (- 2,795 from average)
11/08: the high was 68,000 (+ 3,211 from average)

These maximum differences are the "deviations" from the mean (average) price.

We'll multiply each deviation against itself, squaring them and giving us...

11/06: 16,719,921
11/07: 7,812,025
11/08: 10,310,521

..then average those values to arrive at 11,614,155.66

That number is our variance. We'll then take the square root of the variance to arrive back at a usable (rounded) number of 3,408. This number is what we use for our volatility value.

It's the "standard deviation" for distribution of the price extremes from mean price.

Statistically this means that if price behaved randomly, ~68% of our furthest-from-average price points would be within 3,408 of the average, ~95% of the points would be within 2 standard deviations (+/-6,816) and ~99.7% would be within 3 standard deviations (+/-10,224).

There are a few other methods of calculating volatility but this one is the most common.

Regardless of which one you use, the most important thing is to keep it consistent and not rely on comparisons between results from different models.

Now. We know that price doesn't behave randomly, so is the calculation useless?

Not at all!

Turns out, the market *approximates* random behavior most of the time, and when it stops doing that, it's a good sign something has changed.

One use of volatility is to ensure that you aren't invalidated by random price movements when betting on non-random movement; you probably don't want your stoploss within 1 or 2 standard deviations of the average price (calculated on the timeframe you're watching).

Conversely, if you're hoping for a limit order fill, volatility can give you an indication as to how likely the near-random movements of the market might be to provide a fill.

Higher-than-normal volatility is often a signature of market tops or bottoms, whereas steady trends tend to be less volatile.

Lower volatility in a sideways market can also be a sign of compression waiting to resolve into a trend.

Last but not least, volatility is an excellent foundation for mean-reversion strategies, which tend to perform better during choppy markets.

Adding these sorts of strategies to your arsenal will have you better-prepared for markets that many discretionary traders struggle with.

addendum/misc

-Vol is commonly measured over 30 days

-This calculation is for realized vol, Implied Volatility (IV) in the options market is very different

-This is not the calculation method I use, formula is more useful when tweaked, especially in hard trending markets